Skip to content

arm64: dts: exynos: google: Add Pixel 7 (panther) - #108

Merged
derekross merged 5 commits into
derekross:omamobilefrom
nesmeck:omamobile-gs201-panther-dts
Oct 28, 2026
Merged

derekross merged 5 commits into
derekross:omamobilefrom
nesmeck:omamobile-gs201-panther-dts

Conversation

@nesmeck

@nesmeck nesmeck commented Oct 28, 2026 •

Copy link
Copy Markdown
Contributor

Board file for the Pixel 7 (panther), the gs201 board between the Pixel 7 Pro and the Pixel 7a. The device side (descriptor, board dtbo, display trigger, storage guard and notes) is derekross/omarchy-mobile#79.

Five commits, in upstream order:

  • dt-bindings: arm: google: the google,gs201-panther compatible;
  • the first-light board: USB (NCM and ACM), the cluster watchdogs, a simple-framebuffer on the buffer ABL leaves DPP0 scanning (0xfac00000, read on the phone), ramoops at stock's address; panther matches its base DTB on soc_id (pkg_mode 0), as cheetah does;
  • UFS, with cheetah's and lynx's fixed always-on VCC;
  • the volume and power keys, cheetah's node (same pins), added enabled under the CONTRIBUTING exception since it passed on this phone;
  • the touch rails, LDO25M and LDO26M under pmic-1, disabled, as on the Pixel Fold. ABL leaves both off, with LDO26M programmed at 3.3 V, above stock's 2.5-3.0 V, so it is fixed at 3.0 V; LDO25M gets no voltage and keeps 1.8 V.

The touchscreen node is not here: it needs the focaltech,ft3658 binding from #109 (the FT3658 driver). Once both are in, one more commit adds the node and enables pmic-1.

Tested on a Pixel 7 (bootloader cloudripper-17.0-15902555, slot B), kernel 7.3.0-rc6-omamobile-push1-g08996b25d807: omamobile at 6beaf6d plus this series, the FT3658 series, and the touch node with pmic-1 enabled on top:

  • both watchdogs inactive at probe; the console and the Hyprland lock screen on the simple-framebuffer; SSH over the NCM gadget;
  • UFS: four LUNs, 55 partitions; the omarchy-mobile storage guard leaves only sda20-24 (slot B's boot images) and sda31 (userdata) writable, and the Arch root boots from userdata;
  • the three keys report press and release; the power key wakes the screen;
  • ACPM 15 channels, 14 usable; the S2MPG12 regulators register, vdd-ts 1.8 V and avdd-ts 3.0 V; with the touch node, both come up with one user each and the lock screen takes a PIN by touch.

Checks: tools/kernel-pr-check --build OUT --merge origin/omamobile passes everything except one dtbs_check line, soc_id: ... is not of type 'object', 'integer', 'array', 'boolean', 'null'. gs201-cheetah and gs201-felix print the same line at the base: google.yaml names /soc_id but doesn't describe it. I left that alone rather than guess what the second cell means; happy to add it if you tell me. checkpatch --strict has no errors (the new-file MAINTAINERS note, and the 119-column bootargs line, as lynx's).

🤖 Generated with Claude Code

The Pixel 7 (panther) is a gs201 (Tensor G2) board like the Pixel 7 Pro,
Pixel Fold and Pixel 7a. Add its compatible to the gs201 board list.

Assisted-by: Claude Opus 5.5
Signed-off-by: Alejandro Garcia <skipper.pelygro@gmail.com>
Add a first-light board file for the Pixel 7: USB (NCM and ACM gadget),
the cluster watchdogs from gs201.dtsi, and a simple-framebuffer on the
buffer ABL leaves the panel scanning out. UFS, the panel, touch and the
GPU stay disabled.

panther has the standard package of gs201: stock reports
ro.boot.hw.soc.pkg_mode 0 and soc_id 0x09855001 in its live tree, so
this DTB matches on soc_id as cheetah's does. ABL accepted it.

ABL leaves DPP0 scanning 0xfac00000 (RDMA_BASEADDR_Y8, read on the
phone) for the 1080x2400 sdc-s6e3fc3 panel; the buffer is reserved and
described as a simple-framebuffer. ramoops uses stock's address and
layout, so a log from this kernel can be read after a warm reboot into
stock on slot A.

The board dtbo for ABL (board_id 0x30406, board_rev 0x10000), the
display trigger and the image packing live in the omarchy-mobile
repository (device/panther).

Tested on a Pixel 7 (bootloader cloudripper-17.0-15902555) from slot
B: the kernel boots, both watchdogs are stopped at probe, the console
shows on the panel, and SSH over the USB NCM gadget works.

Assisted-by: Claude Opus 5.5
Signed-off-by: Alejandro Garcia <skipper.pelygro@gmail.com>
Enable the UFS host and its PHY. VCC is the fixed always-on regulator
cheetah and lynx use: stock drives it from gpp0-1 (boot-on), and the
bootloader leaves it on.

omamobile.ufs-probe tells the omarchy-mobile initramfs to load the UFS
drivers and apply its storage guard, which leaves only slot B's boot
images and userdata writable.

Tested on a Pixel 7 from slot B: four LUNs and 55 partitions enumerate,
with the same names and numbers as stock's by-name map, and the guard
leaves sda20-24 (slot B's boot images) and sda31 (userdata) read-write.
The Arch root on userdata boots from it.

Assisted-by: Claude Opus 5.5
Signed-off-by: Alejandro Garcia <skipper.pelygro@gmail.com>
Add gpio-keys for volume down (gpa10-1), volume up (gpa8-5) and power
(gpa10-0, the S2MPG12 PWRONB line, also wired to the SoC). The node is
cheetah's, with the same pins and pin states, which Google's
gs201-common.dtsi gives every gs201 board. It is added enabled, as the
Pixel Fold's volume keys were, since it passed on this phone.

Tested on a Pixel 7 from slot B: each key reports press and release on
the gpio-keys input device, and the power key wakes the screen.

Assisted-by: Claude Opus 5.5
Signed-off-by: Alejandro Garcia <skipper.pelygro@gmail.com>
LDO25M and LDO26M on the main PMIC (S2MPG12) are the supplies of the
FocalTech FT3658 touch controller, its VDD and AVDD; nothing else on the
stock board uses them. Stock constrains them to 1.6-1.95 V and
2.5-3.0 V. Describe those two rails under pmic-1, as the Pixel Fold
does. The sub-PMIC is not described. Leave the node disabled: it is
switched on with the touch controller.

ABL leaves both rails off, LDO25M programmed at 1.8 V and LDO26M at
3.3 V, above stock's range. LDO25M gets no voltage, so the PMIC keeps
its 1.8 V. LDO26M is fixed at 3.0 V, the top of stock's range: stock's
regulator core moves it there before the touch driver switches it on,
and this tree's does the same while the rail is still off.

Tested on a Pixel 7 from slot B, with the node enabled: the S2MPG12
regulators register; vdd-ts reads 1.8 V and avdd-ts 3.0 V, both off,
and no other LDO changes. With the touch controller described as well
(not in this series), both rails come up in order and the controller
answers.

Assisted-by: Claude Opus 5.5
Signed-off-by: Alejandro Garcia <skipper.pelygro@gmail.com>
@derekross

Copy link
Copy Markdown
Owner

Thank you, @nesmeck. This is a clean first board: everything new is in the board file, the five commits are in upstream order, and each one says what ran on the phone.

I checked it the way I check every board on the Pixel 7 Pro's SoC:

  • The Pixel 7 Pro, Pixel Fold and Pixel 7a device trees build byte-identical with and without this series, so nothing changes for the other gs201 phones.
  • The splash buffer (0xfac00000, 1080x2400x4) and ramoops at 0xfd3ff000 use the 12-byte entries gs201 needs. Neither overlaps the SoC's reserved regions (the abl region ends at 0xfa800000, debug at 0xfd3fe000, and bldr-log starts at 0xfd800000).
  • The keys are the Pixel 7 Pro's node with the same pins, so adding them enabled fits the exception in CONTRIBUTING. pmic-1 and the touch rails stay disabled.
  • The 3.0 V on LDO26M matches Google's own panther device tree: gs201-panther-pmic.dtsi in kernel/devices/google/pantah sets regulator-max-microvolt = <3000000> on m_ldo26_reg ("Panther touch controller requires 3.0v typ"). It would be worth citing that file in the touch-rails commit message.
  • Your reading of the soc_id dtbs_check message is right: the Pixel 7 Pro and the Fold print the same line on the base, so it's not this series' problem. Leave it as it is.

Nothing blocks the merge. Two small things, if you push again anyway (a follow-up is fine too):

  1. mem_sleep_default=s2idle in the bootargs, as the Pixel 7 Pro has. Linux has no gs201 deep-sleep sequence, so if anything picks deep (Omarchy's Suspend does by default), the phone never wakes. The device package needs the matching sleep.conf.d drop-in; I've mentioned it on clk: samsung: zumapro: Add the HSI0 USI3 clocks #79.
  2. LDO25M: the Fold describes it as 1.8 V min = max. Doing the same here would pin the voltage stock uses, rather than leaving it at whatever the PMIC holds. Neither choice writes anything today, since the rail is already at 1.8 V.

CI hasn't run yet because first-time contributors' runs wait for my approval. I'll approve it and merge once it's green. #109 merges cleanly alongside this, in either order, and the touch node can follow once both are in.

@derekross
derekross merged commit 04e2dff into derekross:omamobile Oct 28, 2026
1 check failed
derekross added a commit to derekross/omarchy-mobile that referenced this pull request Oct 29, 2026
The matrix after the merges since #118, with each cell going by the
merged notes:

- Pixel 7 (panther): the board file (derekross/linux-omarchy-mobile#108) and the
  FT3658 touch driver (derekross/linux-omarchy-mobile#109) have merged
  into the published kernel, so Boot, Storage, Buttons and USB net say
  "in the published kernel" instead of "in review". Touch stays "no":
  the driver and the touch rails are published, but the board file has
  no touch node, and panther's notes end at boot 8 with the controller
  not answering. BRINGUP-M1's status and NOTES' header say the same.
- Pixel 6 Pro (raven), Charging: no -> works. device-raven 0.1.26 (#112,
  #126) loads the MAX77759 charger at boot with its temperature guard;
  USB-PLAN section 4 records C0-C5 passed on 2026-10-28, a full charge to
  100 % included. The battery node is derekross/linux-omarchy-mobile#112;
  the runs were on an unpublished integration kernel, and the cell says so.
- raven, Torch: still works; the LED's node is now in the published
  kernel (derekross/linux-omarchy-mobile#113), tested on
  omamobile-gs101-torch and not yet run from the published kernel.
- Pixel 9a (tegu), Fingerprint: no -> WIP. FINGERPRINT-PLAN records F0
  and F2-F6b passed on an unpublished kernel: Trusty's TA runs and the HAL
  sees the sensor; enrollment waits for the factory calibration (F6c).
  omamobile-fingerprint 0.1.10 (#101) maps Trusty's stores for it.

Left as they were: tegu's Charging (the MAX77779 charger driver,
derekross/linux-omarchy-mobile#92, is published, but tegu's notes record
no run of it), the gs201 HSI2 pin controller
(derekross/linux-omarchy-mobile#123, no cell), felix (#35 is a plan) and
cheetah (nothing merged for it).

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants